This is a proposal for a randomized trial designed to study the long- term chemopreventive effects of low-dose 13-cis retinoic acid (13-cRA) and of the relatively non-toxic micronutrients beta-carotene and retinol in suppressing premalignant oral lesions (oral leukoplakia). The study's primary hypothesis--i.e., that daily supplementation with oral beta-carotene and retinol will be as effective or more effective than low-dose 13-cRA in producing and maintaining remission in oral leukoplakia--will be tested through assessment of response in the two groups using standard clinical and histologic criteria and measurement of micronuclei frequency. The expression and modulation of a series of biologic biomarkers will be tested through careful correlations between marker expression and modulation and outcome in the two treatment groups. We also plan to prospectively assess and biochemically validate (by assaying for cotinine, retinyl palmitate, and beta-carotene serum levels) the roles of tobacco cessation and dietary intake in the natural history or oral leukoplakia. Preclinical, epidemiologic and early clinical data support the role of retinoids and carotenoids in suppressing upper aerodigestive tract (UADT) and lung carcinogenesis. This proposed long-term, phase III trial has evolved directly from our other two studies of 13-cRA in oral leukoplakia. The first study, now concluded, was in the form of a prospective, randomized trial and established the significant activity of 13-cRA. Preliminary data from the second study indicate the efficacy of low-dose 13-cRA (0.5 mg/kg/d) as short-term (9 month) maintenance therapy and the feasibility of analyzing a limited panel of biomarkers from small tissue samples, which is critical in chemoprevention trials. The proposed study will be significant even if it establishes only that the relatively nontoxic natural agents being tested are effective chemopreventive drugs against oral leukoplakia, as this would imply that they have high potential for chemoprevention of other UADT and lung epithelial cancers. Equally significant, however, will be progress in the laboratory toward the identification of biologic markers as indicators of intermediate endpoints within the multistep carcinogenic process.